Pim1 promotes human prostate cancer cell tumorigenicity and c-MYC transcriptional activity

Cell lines and cell culture

Cell lines were obtained from American Type Culture Collection. Vector control, Pim1 or kinase dead mutant Pim1 (K67M)-overexpressing cells were generated as described [22]. pBabe-Puro-MYC-ER plasmid (gift from Dr. Gerard Evan, University of California at San Francisco, CA, USA) was used to generate retroviruses and infect RWPE1-Neo and RWPE1-Pim1 cells to generate RWPE1-Neo/MYC-ER and RWPE1-Pim1/MYC-ER cells and the cells were maintained as described [23]. To activate c-MYC in chimeric MYC-ER protein, 100 nM of 4-hydroxytamoxifen (4OHT) in ethanol was added to the cells. LNCaP and DU145 cells were maintained in RPMI with 10% fetal bovine serum.

Western blot analyses

Western blotting was performed as described [22] using following antibodies: anti-Pim1 (mouse, 1:500, Santa Cruz), anti-beta-Actin (goat, 1:1000, Santa Cruz), anti-phospho-p21 (rabbit, 1:1000, Santa Cruz), anti-p21 (mouse, 1:1000, Santa Cruz), anti-cyclin E (rabbit, 1:1000, Santa Cruz), anti-androgen receptor (rabbit, 1:500, Santa Cruz) and anti-c-Myc (mouse, 1:500, Santa Cruz).

Cell cycle analysis

Cell cycle was analyzed by fluorescence-activated cell sorting (FACS) as described [24].

MTT cell proliferation assay

MTT Cell Proliferation Assay Kit was purchased from ATCC and assay was performed in accordance with the manufacturer's protocol. Individual absorbance was measured with plate photometer (Bio-tek).

Soft agar assays

These were performed as described [25]. 100 × 10³ and 20 × 10³ of LNCaP and DU145 cells were used, respectively and total number of colonies (≥ cutoff sizes) was counted. For c-Myc inhibitor (10058-F4) treatment, 50 × 10³ of LNCaP and DU145 were used and 100 uM 10058-F4 in 0.25% DMSO or 0.25% DMSO alone was added. Three random low power view-fields were selected and each number of colonies (≥ cutoff sizes) was added.

Xenografts in nude mice

Male nude (nu/nu) mice were obtained from Charles River laboratories. 3 × 10⁶, 3 × 10⁶ and 5 × 10⁶ of RWPE1, LNCaP and DU145 cells were mixed with 200 μl of Matrigel (BD Biosciences), respectively. Cells were injected subcutaneously in both flanks of nude mice. Tumor volumes and cross-sectional areas were calculated as described [26, 27]. Animal care and experiments were carried out according to the protocols approved by the Institutional Animal Care and Use Committees at Vanderbilt University.

Histology and immunohistochemical analyses

Nude mice were sacrificed after 30-38 (RWPE1) or 6-8 (LNCaP and DU145) weeks. Xenografted lesions were taken, photographed, fixed in formalin (10%) and embedded in paraffin for subsequent histology as described [28]. Immunohistochemical analyses were performed as described [28] using following antibodies: activated Caspase 3 (rabbit, 1:500, Cell Signaling), Ki67 (rabbit, 1:50, abcam) and phospho-Histone H3 (rabbit, 1:500, Upstate).

Analysis of androgen-dependent proliferation

15,000 LNCaP cells were plated on 24 well plates. The next day, cells were washed with phosphate-buffered saline (PBS) and phenol red-free RPMI media with charcoal-striped serum was added for androgen starvation. After 2-day starvation, DHT (5α-Dihydrotestosterone) or carrier (ethanol) was added. MTT assays were performed from next day (Day 1). MTT values of DHT-treated cells were divided by those of carrier-untreated cells and triplicate data per group were analyzed.

DHT treatment and quantitative RT-PCR

For DHT treatment, LNCaP cells were washed with phosphate-buffered saline (PBS) and phenol red-free RPMI media with charcoal-striped serum was added for androgen starvation. After 2-day starvation, DHT (0.1, 1, 10 and 100 nM) or carrier (ethanol) was added and the cells were incubated for 2 days. Total RNA isolation, reverse transcription and subsequent quantitative PCR were performed as described [24]. The following primers were used: Prostate specific antigen (PSA) forward (5'-CAA CCC TGG ACC TCA CAC CTA-3'), PSA reverse (5'-GGA AAT GAC CAG GCC AAG AC-3'); human GAPDH forward (5'-ATG GAA ATC CCA TCA CCA TCT T-3'), human GAPDH reverse (5'-CGC CCC ACT TGA TTT TGG-3'); LAMC2 forward (5'-GGA TGA GAA TCC TGA CAT TGA GTG T-3'), LAMC2 reverse (5'-GTC GTG CGG ATC GTT GTA GA-3'); MT1F forward (5'-ACC TGC CCC ACT GCT TCT T-3'), MT1F reverse (5'-TTG CAA GCC GAG GAG AGA CT-3'); UPP1 forward (5'-TCT GGA GGC AGC CTA TGC A-3'), UPP1 reverse (5'-GCA AAC ACC GAG GAC TCC AT-3'); CDKN1C forward (5'-GCC TCT GAT CTC CGA TTT CTT C-3'), CDKN1C reverse (5'-CAT CGC CCG ACG ACT TCT-3'); CUL3 forward (5'-AGA TTT TGA GGC TCC TTT TTT GG-3'), CUL3 reverse (5'-AAA ATT TCT GGC TTT CCA TCT GAA-3'); SOD2 forward (5'-GCT GCA CCA CAG CAA GCA-3'), SOD2 reverse (5'-TCG GTG ACG TTC AGG TTG TTC-3'); VAV3 forward (5'-CTG GTG AAC AAG GGA CAC TCA A-3'), VAV3 reverse (5'-TTT AGG AGT TCT TCG CAG TCC ATT-3'); mouse Pim1 (5'-ATT CCG TTT GAG CAC GAT GAA-3'), mouse Pim1 reverse (5'-TGA AGA GAC AGT TTG CCT GAA GAA-3'). Each mRNA expression was normalized with GAPDH expression.

Luciferase assay

300,000 RWPE1 cells were plated on 12 well plates and grown without supplements (bovine pituitary extract and epidermal growth factor) for 24 hours. Transient transfection was performed with the following plasmids: c-Myc-responsive 4× E-box reporter (gift from Dr. Stephen Hann, Vanderbilt University, TN, USA), pMSCV-MYC and pMSCV empty vector. After 30 hours, cell lysates were harvested and luciferase activity was measured using luciferase assay system (Promega). Luciferase activity was normalized to protein concentration and triplicate data per group were analyzed.

c-Myc inhibitor treatment

c-Myc inhibitor (10058-F4) was purchased from Sigma. Cells were plated on 6-well plates and 50, 100 and 200 uM 10058-F4 in 0.5% DMSO or 0.5% DMSO alone was added on next day when cells were sub-confluent. Cells were harvested after 20 hours and RT-PCR and western blotting were performed from the isolated mRNA and cell lysates, respectively.

c-Myc gene silencing with small hairpin RNA (shRNAmir)

Control and Pim1-expressing LNCaP cells were plated on 60 mm dishes. Lentiviral shRNAmirs (Open Biosystems) against c-MYC and pGIPZ control plasmid were transiently transfected. Transfection efficiency was monitored by GFP fluorescence. After 3 days, cells were harvested, and mRNA isolation followed by reverse transcription and qRT-PCR and western blotting were performed.

Genechip analysis

Total RNA was isolated from RWPE1-Neo/MYC-ER and RWPE1-Pim1/MYC-ER cells with and without 24 hr 4OHT treatment. Genechip analysis was performed in duplicates according to manufacturer's protocol (Affymetrix) using U133A chips. Genes whose expressions were altered at least 1.4-fold with significance (P < 0.005) in pairwise comparisons were identified.

Statistical analyses

Each group was compared using t-test. Values are considered statistically significant at P < 0.05. Quantitative variables are expressed as means + standard deviation while categorical variables are expressed as numbers (%).

Generation of human prostate cell lines at different disease stages, with stable Pim1 overexpression

To examine the effects of Pim1 overexpression on prostate tumorigenesis, we selected three human prostate cell lines at different cancer disease stages: RWPE1, an immortalized, benign, androgen-responsive human prostatic epithelial cell line [29]; LNCaP, a tumorigenic, androgen-responsive human prostate cancer cell line [30, 31]; and DU145, a tumorigenic androgen-independent human prostate cancer cell line [32–34]. We stably expressed Pim1 in all the cell lines and established pools of Pim-1 expressing and vector control (Neo) cells. LNCaP and DU145 cells express high levels of c-MYC endogenously (Figure 1A) but c-MYC levels in the RWPE1 cells were not detectable by western blot (see Figure 2B). To examine Pim1 activity, we assessed a known phosphorylation substrate of Pim1, p21 using a phospho-specific antibody [3]. Phospho-p21 levels were increased 4-fold in RWPE1-Pim1 cells (Figure 1A), showing increased kinase activity of Pim1.

Pim1 has been reported to induce polyploidy and chromosomal instability (CIN) in a passage-dependent manner [22, 24] and the resultant polyploidy/CIN was shown to increase in vitro and in vivo tumorigenicity [25]. We therefore examined the cell cycle profiles of the cells we used in the tumorigenicity assays in the current study to ensure that we utilize cells that have not progressed to polyploidy. As shown in Figure 1B, there were no differences in ploidy between control and Pim1-expressing cells and this is probably due to relatively early passage number (RWPE1, passage 18; LNCaP, passage 27; DU145, passage 23).

Pim1 promotes proliferation and attenuates apoptosis of RWPE1 cells in vivo without malignant conversion

In vitro cell proliferation (MTT) assay showed that there was no discernible change in cellular proliferation due to Pim1 expression in RWPE1 cells (Figure 3A). In addition, RWPE1-Pim1 cells failed to form colonies in soft agar assay (data not shown), verifying that in vitro tumorigenicity of Pim1-expressing RWPE1 cells is due to polyploidy cells driven by chromosomal instability as shown previously [25]. When control and Pim1-expressing cells were grafted in nu/nu nude mice, no tumors formed. H&E stain however showed increased cellularity in Pim1-expressing RWPE1 cells (Figure 3B) and immunohistochemical analysis for Ki67 confirmed elevated proliferation in the tissues of this group (Figure 3C). Apoptosis was also modestly reduced in the Pim1-expressing RWPE1 xenografts as shown by immunostaining for activated Caspase 3 (Figure 3C). These results indicate that although Pim1 could enhance proliferation and suppress apoptosis of RWPE1 cells grown as xenografts, it is insufficient to convert these cells to malignancy.

Pim1 promotes in vitro and in vivotumorigenicity of LNCaP and DU145 cells

We next asked whether Pim1 can enhance the tumorigenicity of established malignant prostate cancer cells. We first tested the androgen-sensitive prostate cancer cell line LNCaP. Pim1 expression increased the soft agar colony formation rate of LNCaP cells by ~2 fold and also led to the formation of larger colonies (Figure 4A). When grafted subcutaneously in nude mice, LNCaP-Pim1 cells formed bigger and heavier tumors (Figure 4B and 4C) with shorter latency (Figure 4D). LNCaP-Pim1 tumors were also more noticeably hemorrhagic by both gross and microscopic examination (Figure 4B and 4E). We analyzed proliferation in the tumors by immunohistochemistry for phospho-histone H3. Proliferation was significantly increased in the LNCaP-Pim1 tumors relative to the LNCaP-Neo control tumors (8.8% ± 1.2 versus 3.5% ± 1.3, P = 0.003). There was also a trend to reduction in apoptosis as determined by staining for activated caspase 3 (0.35% ± 0.25 in LNCaP-Pim1 tumors versus 0.57% ± 0.58 in LNCaP-Neo tumors). Thus Pim1 can promote the tumorigenicity of LNCaP cells. To extend our findings we also used DU145 cells expressing Pim1 in soft agar assays. Pim1 expression increased DU145 cell colony forming potential up to ~6 fold as well as colony size (Figure 5A). When injected into nude mice, DU145-Pim1 cells also formed larger tumors with a shorter latency (Figure 5B,C and 5D). Thus Pim1 expression can further enhance the tumorigenicity of established tumor cells.

Androgen receptor signaling does not affect in vitrocellular proliferation but its transcriptional activity is induced by Pim1

Pim1 has been reported to modulate androgen receptor signaling in prostate cancer cells [35, 36]. To examine whether androgen receptor signaling affects Pim1 tumorigenic function, we compared androgen-dependent cell proliferation rates between control and Pim1-expressing LNCaP cells. As shown in Figure 6A, expression of Pim1 or a Pim1 kinase-dead mutant (K67M) did not affect androgen-stimulated cell proliferation. AR signaling effects on in vitro cell proliferation could be dissociated from other AR signaling functions. Therefore next, we tested androgen receptor transcriptional activity by examining induction of PSA mRNA expression after 5α-Dihydrotestosterone (DHT) treatment. LNCaP-Pim1 cells responded to DHT with significantly higher induction of PSA compared to LNCaP-Neo control cells (Figure 6B). Notably, LNCaP-K67M cells which express the Pim1 kinase-dead mutant showed even lower PSA induction compared to LNCaP-Neo cells (Figure 6B), consistent with the interpretation that the K67M protein functions as a dominant negative mutant. Western blot analysis demonstrated that both Pim1- and K67M-expressing LNCaP cells displayed induction of AR protein expression compared to control LNCaP-Neo cells (2.5 fold and 1.7 fold, respectively) (Figure 6C). Therefore, elevated AR expression can partially explain the increase in PSA levels in LNCaP-Pim1 cells. Other mechanisms are probably operative however, since LNCaP-K67M cells have also increased AR levels but lower PSA expression than control LNCaP-Neo cells (Figure 6B and 6C). These results suggest that Pim1 could promote prostate tumorigenesis by enhancing AR transcriptional activity.

Pim1 enhances c-MYC transcriptional activity

Pim1 tumorigenic function has been intimately linked to c-Myc in several systems. Pim1 is thought to cooperate with c-Myc to promote tumorigenicity by increasing c-Myc protein expression [21, 37] or stimulating the binding of RNA polymerase II leading to increase the transcription of c-Myc target genes [38]. To assess c-MYC transcriptional activity in Pim1-expressing prostate cells we used a c-Myc-responsive 4x-Ebox reporter in luciferase assays. RWPE1-Pim1 cells demonstrated higher c-MYC reporter activity compared to RWPE1-Neo cells (Figure 2A). Furthermore, activity of the c-Myc reporter was suppressed in RWPE1-K67M cells consistent with a dominant negative function of the kinase-dead mutant Pim1 (K67M). This dominant negative action of the K67M mutant is again evident as repression of expression of Cyclin E, a known c-MYC target gene, in RWPE1-K67M cells (Figure 2B).

To obtain a more global view of the possible ability of Pim1 to enhance c-MYC transcriptional activity, we first used an inducible Myc system (MYC-ER, in which c-MYC is fused to a mutant estrogen receptor that responds to 4-hydroxytamoxifen) to identify Myc responsive target genes in RWPE1 prostate cells. We generated stable RWPE1-Neo/MYC-ER and RWPE1-Pim1/MYC-ER cells (Figure 7A). Affymetrix genechip profiling after 24-hr 4-hydroxytamoxifen (4OHT) induction identified 129 c-Myc target genes that were up-regulated or down-regulated in RWPE1-Neo/MYC-ER cells. We then compared these with the genes whose expression is altered by Pim1 expression in the RWPE1-Pim1/MYC-ER (with vehicle treatment) cells. A considerable portion (53 genes, 41%) of the 129 Myc target genes was also altered by Pim1 expression (Table 1). In addition, mRNA expression of some Pim1/MYC target genes in Table 1 was also confirmed in LNCaP and DU145 cells by RT-PCR (Figure 7B) although as expected, there was some variability probably due to the different genetic contexts of these malignant cell lines. LAMC2, MT1F and UPP1, for example, were up-regulated in Pim1-expressing LNCaP cells and genes like CDKN1C, CUL3, SOD2 and VAV3 were repressed in Pim1-expressing DU145 and/or LNCaP cells. These results strongly support the notion that Pim1 cooperates with c-Myc in prostate cancer at least in part by modulating c-Myc-transcriptional activity.

c-MYC inhibition by 10058-F4 treatment or RNA interference abrogates in vitro cellular growth and alters target gene expression of Pim1-expressing prostate cancer cells

To examine if increased c-MYC activity truly contributes to promote tumorigenicity in Pim1-expressing cells, we inhibited c-MYC activity with 10058-F4, a small molecule c-Myc inhibitor. 10058-F4 is known to inhibit hetero-dimerization between c-Myc and Max, so c-Myc is no longer able to trans-activate its transcriptional target genes [39–41]. In soft agar assays, 10058-F4 dramatically suppressed colony formation of LNCaP-Pim1 and DU145-Pim1 cells (Figure 8A). Next, we treated cells with different doses of the inhibitor (0, 50, 100 and 200 uM). Interestingly, 10058-F4 inhibited c-MYC expression itself in LNCaP cells as shown previously but not in DU145 cells (Figure 8B). This phenomenon, where 10058-F4 treatment reduces MYC protein levels in some cell types but not others, even though it inhibits MYC activity in both types of cells, has been noted previously [40, 41].

Remarkably, 10058-F4 also repressed both the endogenous human and transgenic murine Pim1 protein expression in both cell lines (Figure 8B). This effect is not due to a global, non-specific effect on protein expression since neither beta-Actin expression used as a loading control in both cells nor MYC in DU145 cells were changed by inhibitor treatment (Figure 8B). To determine if 10058-F4 treatment affects Pim1 expression at the protein or mRNA levels, we performed RT-PCR to detect the transfected murine Pim1 [24] in LNCaP and DU145 cells. Pim1 mRNA levels were not dramatically changed after inhibitor treatment (Figure 8B), suggesting that 10058-F4 may inhibit Pim1 protein levels via post-translational regulation in addition to c-MYC transcriptional activity.

This novel function of c-MYC inhibitor controlling Pim1 protein expression can not let us examine if decreased tumorigenicity by c-MYC inhibitor is due to inhibited c-MYC activity, repressed Pim1 expression or both. Therefore, to prove that Pim1-induced tumorigenicity is through c-MYC activity, it is necessary to selectively inhibit c-MYC using small hairpin RNA (shRNA). As shown in Figure 8, depending on the levels of c-MYC repression, cells with ~50% c-MYC knock-down (sh1) displayed significant reversal in target gene expression such as LAMC2 and VAV3, but cells with ~25% c-MYC knock-down (sh2) only showed relatively minor change (Figure 8C and 8D). Overall, these results support the notion that MYC is an important mediator for some of the Pim1 effects observed in this model.